Dead PanelDue 7I

A Former User

I have been able to verify that at least one regulator is supplying 3.3 volts, there is another one furthers stepping it down to 1.something. But that's about it.

The micocontroller that is refusing to give you a serial port only needs 3.3v. 1.8v would be for the lcd controller but would need a working microcontroller to evaluate it if also needs fixing.

MikeD

@bearer

Yeah, based on not much more than a gut feeling I think it is going to be one of the micro controllers on the board.

dc42

If the PanelDue won't show up as the Bossa port after you have done all the usual things, then it sounds that either the ATSAM4S4B processor has failed, or the USB port isn't making contact, or there is no +3.3V supply to the ATSAM processor.

The 3.3V supply comes from U4. There should be +3.3V on the end of CA9 closest to the ATSAM processor UA1. The case of the SD card socket can be used as ground when measuring this. There should be 1.2V on LA2. The ATSAM should not be getting hot.

The flashing from the display sounds odd, almost as though the backlight inverter has failed or there is a poor power contact. Are you sure that the backlight is on continuously? Even with the screen dark, I can see it at the edge of the display where the Kapton cables come out of it.

MikeD

I suppose I could try a few more different cables, have been using the one that came with the Duet Wifi boards.
Not really sure how to fully eliminate the USB port not making contact possibility. Plugging usb in does turn on the backlight, so at least GND, and 5V are getting through.

I will verify again the correct voltages later today. I was able to measure +3.3V yesterday in the spot I think you care referring to.

Did also measure 1.X voltage on a regulator, I forgot the exact number, will confirm later also.

Perhaps I described flashing incorrectly:
The backlight is on, always on, cant make it go off even with pressing erase for 1 second followed by reset.
What does, or did initially flash, is the screen itself. Every so often, I would say a 40 second average you would get a quick blip of the screen, kind of like the initial very short white illumination I normally get when powering on, just before the logo displays. Almost as if periodically the unit is trying to boot.

Ironically, I had a similar situation when I initially purchased the screen, before there were any motors or much else hooked up to it. When i was just familiarizing myself with the Duet Wifi and PanelDue I had just those two hooked together.
I was using the 10 pin ribbon cable at the time. Initially I had it hooked to the correct port via the ribbon cable, it came on and life was great.
Throughout the build process at one point I powered on an the screen did not boot. I believe that I had the 10 pin ribbon cable in the wrong port. Powered down, switched to the right port, and I think all was back to normal. Hard to remember, but what I do know is that it did not always boot the screen after that. Sometimes I had to boot a number of times, and I think I am correct in saying that sometimes the screen would boot after being powered for a while. Either way, that is when I initially got familiar with the process of flashing the firmware, bossa etc.
Was able to follow the guides, usb port came up at that time just as it should have, flashed without issue, booted without issue, switched over to 4 pin wire only since ribbon wasn't long enough to reach my desired location.
I put it in its case, and haven't touched it since. Have had no issues for the last 2 months, probably powered down and booted up 30 times since then.
About 2 weeks ago I did have an issue with the screen, or entire system rather going idle during a print. Screen would show IDLE on the corner while machine was still printing.
I originally thought it was on long prints, what would happen is that the screen would essentially stop getting new information, layer number would stop updating, etc.
The machine continued on, and the prints finished as expected. I do not know if pressing the stop button on the screen would have done anything at this point.
During this time the web interface could not connect either, I believe wifi was going idle.
Reboot would fix everything.
At the same time I was also having issues with Fusion 360 taking forever to launch, so I think all of it was really an issue caused by my router. Once I powered cycled it all was better, no more disconnection issues.
Don't know if much of that is helpful or not.
I was surprised that even if it is due to the router and wifi disconnecting that the screen would stop getting information and go idle.
So maybe there were some underlying issues, perhaps from once plugging the 10 pin ribbon into the wrong port.
Once i get my new screen I will be curious to see what happens if I turn off wifi router mid print.

Mike

dc42

Yes, try more USB cables. Some micro USB cables carry power only, not data. Also, connect the USB cable to a PC, not a laptop, because some laptops can't provide enough power. [But if you are getting 3.3V then it should be OK.]

If you haven't already, disconnect the 4-pin cable to the Duet while you are using the USB port.

MikeD

Will do.
I have brand new in the box Duet Wifi for the next project, the current cable I am using came out of that box to eliminate the possibility of of a bad cable.

Danal

I am going to hijack this for about two seconds to say what a remarkable person @dc42 is:

He possess a combination of energy, enthusiasm, patience, technical expertise, and just plain helpfulness that is quite rare. The detailed posts about chips and traces on the board, all to help a single person with a product that's already been replaced, that is what brought all of this to mind; this forum is full of other examples.

Thanks from the entire community, Dave!

Back to your regularly scheduled programming...

MikeD

@dc42
Tried other cables, no difference.
Verified cable functionality by plugging up to an old Tom Tom, detected right away, so comms are goood.

U4 is putting out 3.26 V.

CA9 does have 3.26 at the end specified.

Either end of LA2 does show 1.19 V.

Nothing is getting hot.

Replacing the Atsam looks to be within my skilset, provided I can get one and it doesnt need extra programming with tools I may not have.
Digikey shows a few in stock, not sure if they would be exact matches.
https://www.digikey.com/products/en?keywords= ATSAM4S4B-AU

A Former User

@MikeD said in Dead PanelDue 7I:

https://www.digikey.com/products/en?keywords=ATSAM4S4B-AU

those are all the same part, just different packaging. in practice you can only order the cut tape one in small qty. (but i think you want the lqfp footprint?)
(something weird with that search link, this ought be the correct part ?)

Danal

@MikeD said in Dead PanelDue 7I:

it doesnt need extra programming with tools I may not have.

The bootloader on that chip series is in ROM, so it should be BOSSA-able with no pre-programming on your part.

dc42

@MikeD said in Dead PanelDue 7I:

@dc42
Tried other cables, no difference.
Verified cable functionality by plugging up to an old Tom Tom, detected right away, so comms are goood.

U4 is putting out 3.26 V.

CA9 does have 3.26 at the end specified.

Either end of LA2 does show 1.19 V.

Nothing is getting hot.

Replacing the Atsam looks to be within my skilset, provided I can get one and it doesnt need extra programming with tools I may not have.
Digikey shows a few in stock, not sure if they would be exact matches.
https://www.digikey.com/products/en?keywords= ATSAM4S4B-AU

Yes that's the correct part.

A Former User

@dc42 said in Dead PanelDue 7I:

Yes that's the correct part.

did you click the link or just read the url?

maybe its just on my system it shows the -UU bga version?
(no i think digi key is discarding the - and the new part number indicate rev A giving hits for BA-UU instead of B-AU? and he should look for BA-AU)

dc42

You are quite right, I forgot there there is an extra A in the part number. This is the right one: https://www.digikey.com/products/en?keywords=atsam4s4ba-au

A Former User

Nasty "feature" from digi-key though!

MikeD

Just to confirm:
https://www.digikey.com/product-detail/en/microchip-technology/ATSAM4S4BA-AU/ATSAM4S4BA-AU-ND/4557082

Is the one to order ?

Thank you

dc42

@MikeD said in Dead PanelDue 7I:

Just to confirm:
https://www.digikey.com/product-detail/en/microchip-technology/ATSAM4S4BA-AU/ATSAM4S4BA-AU-ND/4557082

Is the one to order ?

Thank you

Yes.

MikeD

Ordered, will see how it goes.

Any recommendations how to go about replacing it ?
I welcome any suggestions.
I will watch some videos etc to better prepare in the meanwhile.

Removal:
I would assume I want to take the PanelDue as far apart as I can, remove the screen etc.
I do have a reflow gun, after taking apart my plan is to try to heat the chip enough to remove by pushing it off to the side? I could see doing this would possibly bridge some gaps.
Or perhaps before heating try to glue something to the top to use as a lifting device ?
I do also have a "solder sucker" and a Hakko FX-888D soldering iron with many small tips.
Should I try individually going around the legs and removing as much solder as possible ?
Also have copper desoldering braid, perhaps that would be the best method before heating with reflow gun ?

Install:
Hmmm, let me remove it first without collateral damage.....

dc42

Unless you have experience with removing this sort of chip using hot air, you might want to look at ChipQuik or similar low melting point solder, and a no-clean flux pen.

To lift the chip off, either use a vacuum pickup device, or glue a piece of wire to the top (which probably isn't workable with hot air but would be with ChipQuik).

Don't be tempted to lever the chip off the board, because that will lift the PCB traces.

Study some of the videos that you will find on YouTube, and practice on an old hard disk drive if you have one.

MikeD

Thank you, great ideas.
Definitely can scrounge up some old electronics around to practice on.

MikeD

I received the new PanelDue I ordered today, came up as it should have.
At least I can be certain that the Wifi board itself does not have anything going on with it that would prevent the PanelDue from doing its job.